Mobai
/
magic-Dependency
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Update RegexSelector.java

master
zhuyue 2017-05-05 10:47:10 +08:00 committed by GitHub
parent 975adf7072
commit 9e1b7ed3f7
1 changed files with 8 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

28
webmagic-core/src/main/java/us/codecraft/webmagic/selector/RegexSelector.java
View File

@ -26,14 +26,16 @@ public class RegexSelector implements Selector {
if (StringUtils.isBlank(regexStr)) { if (StringUtils.isBlank(regexStr)) {
throw new IllegalArgumentException("regex must not be empty"); throw new IllegalArgumentException("regex must not be empty");
} }
// Check bracket for regex group. Add default group 1 if there is no group.
// Only check if there exists the valid left parenthesis, leave regexp validation for Pattern.
if ( ! hasGroup(regexStr) ){
regexStr = "(" + regexStr + ")";
}
this.regexStr = regexStr;
try { try {
regex = Pattern.compile(regexStr, Pattern.DOTALL | Pattern.CASE_INSENSITIVE); regex = Pattern.compile(regexStr, Pattern.DOTALL | Pattern.CASE_INSENSITIVE);
// Check bracket for regex group. Add default group 1 if there is no group.
// Only check if there exists the valid left parenthesis, leave regexp validation for Pattern.
if ( regex.matcher("").groupCount() == 0 ){
regexStr = "(" + regexStr + ")";
regex = Pattern.compile(regexStr, Pattern.DOTALL | Pattern.CASE_INSENSITIVE);
}
this.regexStr = regexStr;
} catch (PatternSyntaxException e) { } catch (PatternSyntaxException e) {
throw new IllegalArgumentException("invalid regex", e); throw new IllegalArgumentException("invalid regex", e);
} }
@ -44,20 +46,6 @@ public class RegexSelector implements Selector {
this(regexStr, 1); this(regexStr, 1);
} }
private boolean hasGroup(String regexStr) {
int x = StringUtils.countMatches(regexStr, "(") - StringUtils.countMatches(regexStr, "\\(");
int a = StringUtils.countMatches(regexStr, "(?:") - StringUtils.countMatches(regexStr, "\\(?:");
int b = StringUtils.countMatches(regexStr, "(?=") - StringUtils.countMatches(regexStr, "\\(?=");
int c = StringUtils.countMatches(regexStr, "(?<") - StringUtils.countMatches(regexStr, "\\(?<");
int d = StringUtils.countMatches(regexStr, "(?!") - StringUtils.countMatches(regexStr, "\\(?!");
int e = StringUtils.countMatches(regexStr, "(?#") - StringUtils.countMatches(regexStr, "\\(?#");
if (x == (a + b + c + d + e)) {
return false;
}
return true;
}
@Override @Override
public String select(String text) { public String select(String text) {
return selectGroup(text).get(group); return selectGroup(text).get(group);